Pro-inflammatory Cytokines Tumor Necrosis Factor (cid:1) and Interleukin-6 and Survival Factor Epidermal Growth Factor Positively Regulate the Murine GSTA4 Enzyme in Hepatocytes*

We hypothesized that glutathione transferases could be induced and may participate to cellular defenses against the oxidative stress occurring during liver regeneration. Here, we evidenced that murine GSTA1 (mGSTA1), A4, Pi, and Mu are up-regulated during mouse liver regeneration, exhibiting a biphasic pattern of induction correlating early G 1 phase and G 1 /S transi- tion of the cell cycle. Using confocal microscopy immunolocalization and subcellular fractionation, mGSTA4 was demonstrated in both mitochondria and cytosol and found preferentially increased in cytosol during liver regeneration. In addition, mGSTA4 was induced in vivo and in cultured hepatocytes by tumor necrosis factor (cid:1) (TNF (cid:1) ), interleukin-6 (IL-6), and epidermal growth factor (EGF), factors that play crucial roles in hepatocyte survival and proliferation during liver regeneration. However, the mitogenic effect of EGF was not responsible for the induction of mGSTA4. In transient transfec-tions, IL-6 and EGF, but not TNF (cid:1)

Quiescent, differentiated hepatocytes are able to re-enter the cell cycle and proliferate to restore the liver mass following liver deficits resulting from surgical removal or caused by chemicals and viruses. Entry into and progression through early G 1 phase of the cell cycle, also called priming (1), are induced by the cytokines tumor necrosis factor ␣ (TNF␣) 1 (2,3) and interleukin-6 (IL-6) (4) and are required for the hepatocytes to fully respond to growth factors (3). TNF␣ binding to its type 1 receptor successively activates NFB (5), IL-6 expression, and STAT3 (6,7) in the early G 1 phase, which constitute a key signaling pathway during hepatocyte proliferation (4,8). The late G 1 progression and commitment to DNA replication are controlled by growth factors (9 -11) including hepatocyte growth factor, transforming growth factor-␣, and epidermal growth factor (EGF) (1) through the activation of the MEK/ ERK pathway (11).
Several lines of evidence indicate that partial hepatectomy (PH) is rapidly followed by an oxidative stress due to increased reactive oxygen species (ROS) and nitric oxide (NO) production leading to lipid peroxidation (12)(13)(14).
Regarding ROS, TNF␣ is now recognized as playing a crucial role in the production of these species during liver regeneration. In support of such a role, it is worth noting that the multimerization of TNF␣ type I receptor following binding of TNF␣ has indeed been shown to lead to recruitment of TRAF (TNF receptor-associated protein), a protein involved in the signaling pathway regulating ROS production in mitochondria (15), probably through the inhibition of complex III of the electron transport chain (16).
Both ROS and NO have been demonstrated to contribute significantly to induce hepatocyte proliferation. Indeed, in transgenic mice with targeted disruption of TNF␣ type I receptor (TNFRI) or of the type II nitric-oxide synthase (iNOS), the enzyme that catalyzes the formation of NO from arginine, hepatocyte proliferation after PH is strongly impaired (2,12). In this context, it then appears that TNF␣-related oxidative stress functions as a signaling pathway rather than eliciting deleterious effects. Such a function is mediated through the activation of redox-sensitive proteins, especially the transcription factor NFB (17), which on one hand allows the occurrence of proliferation by transactivating cell cycle genes such as c-fos and c-jun controlling the G 0 /G 1 transition, and on the other hand, attenuates deleterious responses resulting from oxidative stress (e.g. activation of caspases by trans-activating the anti-apoptotic gene Bcl XL that prevents activation of caspases (18)).
NFB also induces the mitochondrial uncoupling protein UCP-2 (13) and iNOS (19), both of which contribute to reduce oxidative stress (1). UCP2, an inner mitochondrial channel for protons, plays a major role in limiting production of ROS (13) by dissipating the electrochemical gradient. NO, whose production is enhanced by induction of iNOS, although it participates in lipid peroxidation (14), also exhibits a cytoprotective effect by preventing apoptosis through S-nitrosylation of caspases, which strongly inhibits caspase activities (20). This last mechanism is most likely involved in liver regeneration because in iNOS knock-out mice, PH is followed by a strongly increased caspase-3 activity and hepatocyte death (12). Along with the TNF␣-dependent activation of antioxidant defenses, it is worth noting that cytokine IL-6 and growth factors have also been shown to favor hepatocyte survival during liver regeneration through stimulation of anti-apoptotic gene products (21,22).
In normal hepatocytes, an excess of ROS is also neutralized through the action of thiols, especially glutathione (GSH), of which the content is increased during liver regeneration (21). In addition, glutathione S-transferases (GST) of the alpha class detoxify organic hydroperoxides and protect cells against oxidative stress (22). For instance, transfection of hGSTA2 in K562 cells results in cell protection toward H 2 O 2 -induced lipid peroxidation (23). The GSTA4 enzyme, the ␣ subunit that exhibits the highest activity against 4-hydroxynonenal (4-HNE) (24,25), also efficiently protects against oxidative damage mediated by this cytotoxic product of lipid peroxidation generated by ROS overproduction (26). In addition, we have recently demonstrated that ROS overproduction induced by hepatic iron overload is correlated with an increase in mGSTA4 expression (27).
These observations suggest that GST enzymes could participate in defenses against oxidative stress during liver regeneration. However, to date, little is known about regulation of GST expression and activity during liver regeneration. Lee and Boyer (28) have reported a decrease in mRNA levels of several GSTs at 12 h post-PH, whereas Mori et al. (29) have shown a higher expression of GSTPi after 2 and 3 days post-PH. To our knowledge, no study has attempted to correlate the expression of GSTs with the oxidative stress occurring during the first hours following PH.
The aim of this paper was, first, to study the expression and activities of several GST subunits, including the mGSTA4 isoform, during the liver regeneration after a two-thirds hepatectomy in mouse. Then, we investigated whether oxidative stress, the pro-inflammatory cytokines TNF␣ and IL-6, and growth/survival factor EGF, which are involved in proliferation and survival of hepatocytes during liver regeneration, may regulate the expression of mGSTA4 both in vivo and in hepatocytes in primary culture.
Animal Experiments-Balb/c male mice (8 weeks old, Janvier laboratories, Le Genest, France) were subjected to 70% partial hepatectomy under ether anesthesia and were killed at various times post-PH. As controls, animals underwent a sham operation consisting of a laparot-omy without tissue resection. L-N 6 -(1-Iminoethyl)lysine dihydrochloride (NIL; Calbiochem) was injected intraperitoneally at a dosage of 40 g/g of body weight in normal and hepatectomized animals 1 h before the surgery. Animals were killed 2 h after NIL injection. TNF␣, IL-6 (Promocell, Heidelberg, Germany), and EGF (Promega, St. Quentin Fallavier, France) were given intraperitoneally in 0.2 ml of sterile pyrogenfree saline (0.9% NaCl) with 0.1% bovine serum albumin at 40, 80, and 125 ng/g of body weight, respectively. Control animals received the corresponding sterile saline vehicle. Animals were killed 2 h after TNF␣ and IL-6 injection and 6 or 24 h after EGF injection.
RNA Extraction and Northern Blot Analysis-Total RNA from cells or liver biopsies were isolated using an SVRNA extraction kit (Qiagen, Valencia, CA). Ten g of total RNA from each sample were used for Northern blot analysis. Blots were hybridized with the corresponding 32 P-labeled cDNA probe ([␣-32 P]dCTP, 3000 Ci/mmol, Amersham Biosciences) at 65°C overnight. Blots were washed at moderate stringency and exposed to radiograph films. The 18 S ribosomal probe was used as control.
Protein Extraction and Western Blot Analysis-Cultured hepatocytes and liver biopsies were homogenized in lysis buffer (HEPES, pH 7.5, 50 mM, NaCl 150 mM, EDTA 1 mM, EGTA 2.5 mM, 0.1% Tween 20, 10% glycerol, ␤-glycerophosphate 10 mM, sodium fluoride 1 mM, sodium orthovanadate 0.1 mM, phenylmethylsulfonyl fluoride 0.1 mM, leupeptin 10 g/ml, and aprotinin 10 g/ml). CDK1 was purified from liver extracts using p9CKS hs1 beads and recovered with sample buffer as described previously (9). Protein concentrations were determined using the Bradford method. Proteins were fractionated by SDS-PAGE (12.5%), transferred to a polyvinylidene difluoride membrane and incubated further for 2 h in PBS containing 3% bovine serum albumin and then overnight with primary antibody. Membranes were washed twice before incubation with secondary antibody for 1 h. Proteins of interest were visualized using the chemiluminescence reagent ECL (Interchim).
Determination of Intracellular MDA and 4-HNE Levels-MDA and 4-HNE levels were determined using lipid peroxidation assay kits (Calbiochem). Liver biopsies were washed in ice-cold 0.9% NaCl and sonicated in 20 mM Tris-HCl, pH 7.4, to ϳ10% (w/v). Liver extracts were centrifuged at 3000 ϫ g for 10 min at 4°C, and the supernatant was collected prior to determination of total protein concentration and the MDA and 4-HNE colorimetric assays.
Mitochondrion Isolation-Liver mitochondria were isolated from freshly harvested livers by differential centrifugation in ice-cold H medium containing 210 mM mannitol, 70 mM sucrose, and 2 mM HEPES buffer (pH 7.4) (32). The purified pellets were suspended in 200 l of buffer containing 150 mM KCl, 0.5 mM malonate, 0.1 mM oxoglutarate, and 10 mM HEPES.
Immunohistochemistry-Mice were anesthetized and perfused through the portal vein with 4% paraformaldehyde in 0.1 M sodium cacodylate for 15 min at a flow rate of 10 ml/min. Tissue fragments were washed in 0.1 M PBS for 4 h and in 10% glycerol-PBS overnight and were frozen in liquid nitrogen-cooled isopentane. Frozen tissue sections were mounted on glass slides coated with 10% gelatin in PBS and incubated in PBS containing 3% bovine serum albumin for 30 min. They were then covered with a solution of anti-rabbit GSTA4 (1/100) or anti-rabbit cytochrome c (1/100) antiserum for 1 h at room temperature. Sections were washed with PBS, incubated for 2 h with goat anti-rabbit IgG conjugated to rhodamine (1/200), washed, and mounted. Serial z-axis optical analysis of sections was done at 1-m intervals using a laser scanning confocal microscope (Confocal Leica TCS NT).
Statistical Analysis-Values were expressed as the mean Ϯ S.D. Student's t test was used for the estimation of statistical significance. A p value less than 0.05 was considered statistically significant.

Induction of Several GST Isoforms during Liver
Regeneration-The expression of the different GST isoforms was analyzed during liver regeneration over a 96-h period after a twothirds hepatectomy of Balb/c mice. To ensure that liver regeneration occurred as previously described, the relative mRNA levels of the two cell cycle genes, cyclin D1 and CDK1, were analyzed. A clear induction of cyclin D1 transcripts was observed between 30 and 96 h, whereas the level of CDK1 mRNA was increased between 48 and 96 h post-PH with a maximal expression at 72 h (Fig. 1A) as previously reported (37).
Levels of mGSTA4, A1, and Pi mRNAs were augmented during liver regeneration exhibiting a biphasic induction of expression (Fig. 1B). An initial increase was observed within 1 h post-PH, and mRNA levels remained high for 12 to 24 h. The second peak of induction took place at 40 -48 h, and the increase in transcript levels was observed until 72-96 h. In sham-operated mice, mGSTA4, A1, and Pi mRNA levels showed a transient and moderate increase between 1 and 8 h post-PH but not thereafter. mGSTMu mRNAs were detected at very low levels, and no significant change was observed during the first 96 h of regeneration. mGSTA4, A1, Pi, and Mu proteins were further investigated by Western blot (Fig. 2A). Expression of mGSTA4, A1, Pi, and Mu proteins exhibited a biphasic pattern as observed for mRNAs. The first induction occurred between 0.5 and 4 h after PH; mGSTA1 was very transiently induced at 0.5 h, whereas up-regulation of mGSTA4, Mu, and Pi remained elevated between 0.5 and 4 h. The second induction of the four GST proteins took place between 24 and 72 h depending upon the GST isoforms. In sham-operated mice, a slight increase of mGSTA4, A1, Mu, and Pi proteins was also detected at 1 and 12 h post-PH as observed with the corresponding mRNAs, which returned to control values thereafter except for mGSTPi, which remained elevated until 48 h.
To confirm the induction of several GSTs, during the first hours post-PH, we measured GST activities using three different substrates: CDNB, a common substrate for mGSTA4, A1, Mu, and Pi; and 4-HNE and EA, specific substrates for GSTA4 and GSTPi, respectively (22). GST activities measured with CDNB, 4-HNE, and EA were all significantly increased between 0.5 and 2 h after PH compared with sham-operated animals. The induction of GST expression observed between 24 and 72 h was also correlated with a strong induction of GST activities using CDNB as a substrate (Fig. 2B).
mGSTA4 Is Located in the Cytosol and Mitochondria of Hepatocytes-The subcellular localization of mGSTA4 was analyzed in normal mouse liver using the indirect immunofluorescence technique and confocal microscopy (Fig. 3A). A punctuated staining typical of a mitochondrial distribution was observed for cytochrome c whereas mGSTA4 exhibited both a punctuated and an intense homogeneous cytosolic staining.
To confirm these results, mGSTA4 protein expression was analyzed by Western blot in the cytosol and mitochondria of normal or regenerating livers after cell fractionation. The reliability of the cell fractionation was demonstrated in cell extracts from normal and 24-h post-PH regenerating livers by performing Western blots of the mitochondrial cytochrome c and cytochrome oxidase proteins, which were detectable mainly in mitochondrial fractions while albumin was observed only in cytosols (Fig. 3B).
By Western blot the mGSTA4 protein was detectable in both mitochondrial and cytosolic fractions from normal liver, regenerating livers 1 and 24 h post-PH, and sham-operated livers 24 h post-laparotomy (Fig. 3C). The signal obtained in cytosols was much higher than in mitochondria and was increased at 1 and 24 h post-PH compared with normal and sham-operated animals. No change in mGSTA4 level was found in mitochondrial extracts during liver regeneration even after a longer exposure time of the blots (data not shown).
TNF␣, IL-6, and EGF, but Not Lipid Peroxidation, Induce mGSTA4 Expression in Vivo-During the first hours post-PH, ROS are produced in mitochondria (13), although NO is released. These compounds contribute to enhanced lipid peroxidation (14).
To determine whether 4-HNE and MDA, two metabolites known to induce GSTA4 (38), could be responsible for the early induction of mGSTA4 following PH, the levels of 4-HNE and MDA, were measured in 1-h regenerating livers, and found to be increased by a 3-fold factor (p Ͻ 0.001) when compared with normal ( Fig. 4A) or sham-operated animals (data not shown). We also measured lipid peroxidation by analyzing 4-HNE/MDA levels and mGSTA4 expression by Western blot in mice injected with the selective inhibitor of iNOS, L-N 6 -(1-Iminoethyl)lysine dihydrochloride, which blocked the rise of NO production and lipid peroxidation in hepatectomized liver (14). The level of lipid peroxidation was strongly decreased in both nonhepatectomized (25-fold; p Ͻ 0.01) and 1-h regenerating (3.9-fold; p Ͻ 0.001) livers of NIL-treated mice when compared with normal and hepatectomized animals that had not been injected with NIL, respectively (Fig. 4A). The levels of mGSTA4 were not affected by NIL treatment in regenerating liver, whereas in normal liver, injection of NIL increased mGSTA4 expression (Fig. 4B). These results therefore favor the conclusion that the increase in 4-HNE and MDA content after PH did not trigger mGSTA4 induction. Thus, we postulated that the cell cycle priming factors, TNF␣ and IL-6, produced within 15 min post-PH, and survival/ growth factor EGF, which cooperates with TNF␣ and IL-6 to induce progression through the cell cycle and favor survival of hepatocytes (1), could be involved in this induction. TNF␣ and IL-6 injections to normal mice induced expression of mGSTA4 but not mGSTA1, used as control, compared with mice injected with the vehicle only (Fig. 4C).
In normal livers at 6 and 24 h after EGF administration, expression of mGSTA4 was significantly induced compared with the levels of mGSTA4 in livers of mice injected with the vehicle and noninjected animals (Fig. 4D). To determine whether EGF also induced proliferation of liver cells, expression of CDK1 protein, a cell cycle marker of S, G 2 , and M phases (9, 10), was studied and compared with its expression in regenerating livers at 40 and 72 h post-PH. In livers of control and EGF-treated mice, CDK1 was not detected, whereas its expression was strongly induced in regenerating livers. These results indicated that TNF␣, IL-6, and EGF up-regulated mGSTA4 expression in normal liver.
TNF␣, IL-6, and EGF Induce mGSTA4 Expression in Primary Cultures of Mouse Hepatocytes-Primary cultures of mouse hepatocytes were used to confirm the up-regulation of mGSTA4 by TNF␣, IL-6, and EGF and analyze the level of regulation.
The addition of TNF␣ and IL-6 to the culture medium at 24 h led to induction of the mGSTA4 protein, whereas the levels of mGSTA1 and albumin, used as controls, were unaffected (Fig.  5A). Stimulation by EGF at 4 or 24 h after plating transiently induced mGSTA4 at 24 and 48 h, respectively, whereas the level of mGSTA1 was not increased after EGF stimulation (Fig.  5B). To verify if stimulation by these factors resulted in an increase in mGSTA4 mRNA levels, a Northern blot analysis using mGSTA4 cDNA was performed on EGF-, TNF␣-, and IL-6-treated hepatocytes. Induction of mGSTA4 mRNAs by EGF, TNF␣, and IL-6 was confirmed (Fig. 5C).
To determine whether the induction of mGSTA4 by these soluble factors occurred at the transcriptional level, transfection experiments were performed using the human GSTA4 promoter, which we had previously isolated and characterized (35). A significant increase in luciferase activities was observed after stimulation with EGF and IL-6, but not with TNF␣, when cells were transfected with the pGL3-GSTA4-(1-1571) plasmid carrying 1571 base pairs of the hGSTA4 promoter upstream of the luciferase reporter gene (Fig. 5D), indicating that induction of hGSTA4 by EGF and IL-6 was most likely due to a transcriptional activation of the GSTA4 gene.
Proliferating and/or Survival Pathways Are Involved in the Induction of mGSTA4 by EGF-To confirm that proliferating and/or survival pathways were involved in the induction of mGSTA4 expression by EGF, we used inhibitors of different signaling pathways, namely Ly 294002, SB 203580, and U0126 which inhibit PI3K, p38 MAPK, and MEK, respectively, and analyzed mGSTA4 expression by Western blot.
Treatments of EGF-stimulated hepatocytes by Ly 294002, SB 203580, and U0126 led to a strong decrease in mGSTA4 protein expression, whereas GSTA1 was not modified (Fig. 6A). Because interference between the PI3K and MEK/ERK pathways has been evidenced in many cell types, the effects of these inhibitors on the phosphorylation of AKT (P-AKT) and ERK1/2 (P-ERK), respective substrates of PI3K and MEK, were investigated by Western blot (Fig. 6A). Ly 294002 was found to strongly diminish P-AKT and P-ERK levels, whereas U0126 almost completely abolished P-ERK and slightly decreased P-AKT. SB 203580 affected neither P-AKT nor P-ERK. In addition, these three molecules strongly inhibited DNA replication in EGF-stimulated hepatocytes (Fig. 6B).
The ability of these inhibitors to down-regulate hGSTA4 transcriptional activity by transient transfection of pGL3-GSTA4-(1-1571) plasmid was investigated. Four hours after transfection, hepatocytes were treated with SB 203580, Ly 294002, and U0126 in the absence or presence of EGF for 24 h. As expected, luciferase activity was enhanced in the presence of EGF. SB 203580 slightly decreased luciferase activities in both nonstimulated and EGF-stimulated cells, whereas treatments by Ly 294002 and U0126 strongly down-regulated reporter gene activity in both nonstimulated and EGF-stimulated cultures (Fig. 6C). DISCUSSION Liver regeneration following partial hepatectomy is associated with an overproduction of ROS, which probably play a critical role in the induction of hepatocyte proliferation (1). Several mechanisms of defense are activated to neutralize this ROS excess (12,13). Among the protective defense systems in hepatocytes, GSTs, particularly GSTA4, are recognized to play an important role in the elimination of lipoperoxidation products in various physiopathological situations (39,40).
Only a few studies have dealt with the regulation of GSTs during liver cell growth without detailed kinetics of GST expression during the first 3 days post-PH (28,29). Our results have clearly demonstrated for the first time that several GSTs belonging to distinct classes, in particular mGSTA4, exhibit a biphasic pattern of induction concomitant with the two critical steps of liver regeneration, i.e. entry into the cell cycle (the so-called priming) and the commitment to DNA synthesis, which are controlled by the pro-inflammatory cytokines TNF␣ and IL-6 and by growth factors, respectively (1).
Oxidative stress, demonstrated by H 2 O 2 production in mitochondria (13) and increased levels of MDA and 4-HNE products in cytoplasm detected in this study, is observed in early G 1 phase during liver regeneration. We postulated that this increase in lipid peroxidation products was associated with an increase in mitochondrial and/or cytosolic mGSTA4 content. Several recent studies have dealt with immunolocalization of GSTA4 in hepatocytes and led to contradictory results. This enzyme has been reported to be located in both mitochondria and cytosol (41), only in mitochondria (42), or predominantly at or near the plasma membrane (43). Our observations based on both immunofluorescence confocal microscopy and subcellular fractionation support a distribution of mGSTA4 in both mitochondria and cytosol and a preferential increase of mGSTA4 content in the cytosol of regenerating hepatocytes. The mitochondrial and cytosolic localization of mGSTA4 is compatible with the formation of ROS in mitochondria and lipid peroxidation-derived aldehydes in membranes that diffuse within the cell and attack targets far from the site of their original production.
The induction of mGSTA4 during the first hour post-PH could result from oxidative stress, because 4-HNE is known to be substrate and inducer of this GST isoform (25,38). However, our results show that, in both normal and regenerating liver, NIL, a specific inhibitor of iNOS, diminished the levels of 4-HNE and MDA, whereas mGSTA4 protein content was unchanged or augmented in regenerating and normal liver, respectively. These findings favor the conclusion that the expression of mGSTA4 is not strictly correlated to the levels of lipid peroxidation and more precisely to the contents of 4-HNE and/or MDA. Nevertheless, it cannot be totally excluded that free radicals rather than lipoperoxidation products may contribute to the up-regulation of mGSTA4 during liver regeneration.
In the regenerating liver, induction through NFB activation (17) of cellular defenses involved in the neutralization of the excess of ROS and lipid peroxidation products, such as the mitochondrial UCP-2 protein (13), iNOS (14), and the antiapoptotic proteins Bcl XL and Akt (44,45), are dependent upon stimulation by the pro-inflammatory cytokines TNF␣ and IL-6 and the growth/survival factors transforming growth factor-␣, hepatocyte growth factor, and EGF. Here, we show that injection of TNF␣, IL-6, and EGF to normal mice resulted in an induction of mGSTA4, whereas mGSTA1 remained unchanged. The biphasic induction of mGSTA4 during the first hours after PH could result from the overproduction of the pro-inflammatory cytokines TNF␣ and IL-6 involved in the cell cycle priming and activation of survival pathways by growth factors such as EGF. This hypothesis is strongly reinforced by our in vitro data, demonstrating that mGSTA4 expression was also increased after stimulation by TNF␣, IL-6, and EGF. The second peak of induction of mGSTA4, occurring between 24 and 48 h post-PH, could be due to the increase in growth factors in plasma that promotes the G 1 /S transition (11). However, our data indicated that mGSTA4 induction by EGF is not necessarily correlated to progression in late G 1 or S phases. Indeed, in vivo injection of EGF induced mGSTA4, whereas hepatocytes did not proliferate, as previously reported (1) and confirmed in this study by the absence of CDK1 expression, a cell cycle marker of S, G 2 , and M phases (9).
Previous studies have demonstrated modulation of GST expression by cytokines and growth factors in the liver. An increase in hGSTA1 and A2 by IL-4 in cultured human hepatocytes (46) and a marked decrease in rGSTA2 and M1 mRNA levels by IL-1␤ in rat hepatocytes (47) have been evidenced. Moreover, rGSTP1 expression is strongly induced by EGF in cultured rat hepatocytes (48). However, this is the first time, to our knowledge, that induction of several GSTs has been evidenced during liver regeneration and that a correlation has been established between this induction and the soluble factors essential for hepatocyte survival and proliferation.
Our results also suggest that IL-6 and EGF positively regulate hGSTA4 expression at the transcriptional level, whereas TNF␣ could act rather at a post-transcriptional level as previ- FIG. 5. Up-regulation of mGSTA4 expression by EGF, IL-6, and TNF␣ in primary cultures of mouse hepatocytes. A, Western blot analysis of mGSTA4, A1, and albumin expression in untreated hepatocytes (CONTROL) or in hepatocytes exposed to IL-6 (20 ng/ml) or TNF␣ (20 ng/ml). Cytokines were added 24 h after cell seeding, and hepatocytes were harvested 6 h after treatments. B, Western blot analysis of mGSTA4 and A1 expression in hepatocytes treated (ϩ) or not (Ϫ) with EGF (50 ng/ml) and analyzed at the indicated times after plating. Stimulation by EGF was performed at either 4 or 24 h after plating. C, Northern blot analysis of mGSTA4 in untreated hepatocytes or in those stimulated by EGF, TNF␣, or IL-6. Treatments were performed for 6 h starting 24 h after plating. These data are representative of three independent experiments. D, Hepatocytes were transfected with plasmids carrying the 1.5-kb proximal promoter (pGL3-GSTA4(-1571)) or deleted the promoter (pGL3- GSTA4(1-165)) of the hGSTA4 upstream luciferase reporter gene. Cells were exposed to EGF for 24 h after a 4-h transfection and to TNF␣ or IL-6 for 6 or 18 h, respectively, after transfection. At 48 h of culture, cells were harvested and lysed for luciferase activity assays. Values are means Ϯ S.D. of three independent experiments. Statistical significance was calculated between control and treated cells transfected with the pGL3-GSTA4(1-1571) plasmid. **, p Ͻ 0.01 with EGF and IL-6 treatments. Inset, negative (pGL3 basic) and positive (pGL3 prom) firefly luciferase controls.
ously shown for IL-1␤ on rGSTA2 and M1 regulation (47) in rat hepatocytes. Another hypothesis can be proposed to explain why TNF␣ did not induce luciferase activity after transfection with the 1.5-kb hGSTA4 sequence promoter. Indeed, in a previous study, we found several putative binding sites for AP1, Sp1, and STAT as well as NFB within the 1.5-kb sequence upstream of the transcription start site of the hGSTA4 gene (35). More recently, we localized several other putative binding sites for transcriptional factors, including NFB, AP1, and CREB (cAMP response element-binding protein), within the 0.5 kb upstream of the promoter region used for our transfection assays (data not shown). Therefore, we cannot rule out that TNF␣ could activate hGSTA4 transcription via the putative NFB binding site located upstream of the 1.5-kb promoter sequence.
The p42/44 ERK1/2 (11), p38 MAPK (49), and PI3K (44) signaling pathways activated during liver regeneration are essential for both the proliferation and survival of hepatocytes during the hepatic regenerative process. Recently, several reports have indicated that MAPK pathways are involved in the regulation of GST enzymes. Indeed, Kang et al. (50) have demonstrated that the activation of p38 MAPK and PI3K during oxidative stress leads to the induction of rGSTA2. In addition, Yin et al. (51) have demonstrated that GSTPi has protective effects against H 2 O 2 -mediated cell death via activation of p38 MAPK, ERK, and NFB and repression of the c-jun N-terminal kinase signaling pathways. Here, we have shown that the use of the specific inhibitors Ly 294002, U0126, and SB 203580 of the protein kinases PI3K, MEK, and p38 MAPK, respectively, prevented the induction of mGSTA4 expression and the transcriptional activation of the hGSTA4 promoter-luciferase construct by EGF.
Altogether, these data demonstrate the induction of several GST isoforms, including mGSTA1, A4, Mu, and Pi, during liver regeneration. They also strongly suggest that the pro-inflammatory cytokines TNF␣ and IL-6 and the growth/survival factor EGF, which control hepatocyte survival and proliferation during liver regeneration, might be involved in the up-regulation of mGSTA4 via PI3K and/or MAPK pathways. Thus, mG-STA4 could be a target gene induced by survival factors and could contribute to cellular defenses against oxidative stress in hepatocyte.